The advent of Industry 4.0 has reshaped modern manufacturing, driven by breakthroughs in cutting-edge technologies. A key example is the widespread deployment of sensors, which capture and transmit large volumes of operational data. This data surge has fueled the development of advanced Artificial Intelligence (AI) applications, enhancing manufacturing intelligence and efficiency. A key enabler of such intelligence is Time-Series Forecasting (TSF), which leverages historical data to predict future trends and events, thereby providing actionable insights for proactive decision-making. In parallel, Digital Twin (DT) technology has gained significant prominence due to its capacity for bidirectional communication with physical manufacturing systems, enabling unprecedented levels of real-time monitoring, control, and optimization. Despite their benefits, the combined adoption of TSF and DT technologies presents considerable challenges, particularly in developing integrated, closed-loop systems. This study addresses this gap by proposing a novel framework that unifies TSF with DTs for the early detection of potential collisions between manufacturing assets. The framework is demonstrated using a robotic assembly line, with a detailed account of the training and deployment process of a TSF–DT pipeline. The proposed proof-of-concept is designed to be generalizable, offering applicability across diverse manufacturing systems.
Kalach et al. (2026) studied this question.